Loading…
Loading…
Fine-tuned Luganda-English translation model optimized for child helpline services. Trained on 42,510 child-welfare conversations with crisis detection and emotion preservation. Achieves 0.616 BLEU score (422% improvement over baseline). Specialized for counseling, family support, and crisis intervention contexts with cultural sensitivity for Ugandan child communication patterns. Includes safety filters and harmful content detection. 8 epochs training on Helsinki-NLP/opus-mt-mul-en base model with custom tokens for crisis/emotional contexts.
Model Name: opus-mt-lg-en-finetuned
Version: 8
Task: Translation
Framework: pytorch
Export Method: transformers
MLflow Experiment: translation-lg-en
Training Date: 2025-08-02 05:14:25
Child helpline real-time translation, crisis intervention support, counseling sessions, family support communications, educational guidance for Ugandan children and adolescents
Best on conversational/emotional content, may struggle with technical terms. 512 token limit. Enhanced emotional sensitivity may over-interpret neutral text. Requires <2s response time for crisis situations.
| Parameter | Value |
|---|---|
| Team Config Notes | Increased Batch by 2, epochs by 4 and max length * 2 |
| Total Parameters | 77518848 |
| Trainable Parameters | 76994560 |
| Vocab Size | 64172 |
| Max Length | 512 |
| Language Pair | lg-en |
| Language Name | Luganda |
| Model Name | Helsinki-NLP/opus-mt-mul-en |
| Dataset Config Primary Dataset | custom |
| Dataset Config Custom Datasets | ['dataset/custom/luganda/converted_data.jsonl', 'dataset/custom/luganda/luganda_parallel.jsonl'] |
| Dataset Config Validation Split | 0.15 |
| Dataset Config Max Samples | None |
| Training Config Learning Rate | 4e-05 |
| Training Config Batch Size | 4 |
| Training Config Num Epochs | 12 |
| Training Config Max Length | 256 |
| Training Config Weight Decay | 0.01 |
| Training Config Warmup Steps | 3000 |
| Evaluation Config Metrics | ['bleu', 'chrf', 'meteor'] |
| Evaluation Config Test Size | 150 |
| Team Config Assigned Developer | Marlon |
| Team Config Priority | medium |
| Metric | Value |
|---|---|
| Train Samples | 108613.0000 |
| Validation Samples | 19167.0000 |
| Total Samples | 127780.0000 |
| Baseline Bleu | 0.1683 |
| Baseline Chrf | 37.6363 |
| Loss | 0.0793 |
| Grad Norm | 4.5997 |
| Learning Rate | 0.0000 |
| Epoch | 11.0000 |
| Eval Loss | 0.2365 |
| Eval Bleu | 0.5028 |
| Eval Chrf | 69.4387 |
| Eval Runtime | 2114.2653 |
| Eval Samples Per Second | 9.0660 |
| Eval Steps Per Second | 2.2670 |
| Bleu Improvement | 0.3345 |
| Bleu Improvement Percent | 198.8019 |
| Chrf Improvement | 31.8024 |
| Chrf Improvement Percent | 84.4992 |
| Train Runtime | 51057.2887 |
| Train Samples Per Second | 25.5270 |
| Train Steps Per Second | 6.3820 |
| Total Flos | 47003705065930752.0000 |
| Train Loss | 0.1499 |
| Final Eval Loss | 0.2365 |
| Final Eval Bleu | 0.5028 |
| Final Eval Chrf | 69.4387 |
| Final Eval Runtime | 2114.2653 |
| Final Eval Samples Per Second | 9.0660 |
| Final Eval Steps Per Second | 2.2670 |
| Final Epoch | 11.0000 |
pip install transformers torch
from transformers import pipeline
# Load the model
translator = pipeline("translation", model="openchs/lg-opus-mt-multi-en-synthetic-v1")
# Translate text
result = translator("Nsobola okuba n’ennamba yo ey’essimu??")
print(result[0]["translation_text"])
212ff762127b4d569dcf850b23f6268b22BITZ AI TEAM
If you use this model, please cite:
@misc{opus_mt_lg_en_finetuned_8,
title={opus-mt-lg-en-finetuned v8},
author={BITZ AI TEAM},
year={2025},
publisher={Hugging Face},
url={https://huggingface.co/marlonbino/opus-mt-lg-en-finetuned}
}
This model card was automatically generated from MLflow metadata and user inputs.